Latest news with #physicalAI
Wall Street Journal
17 hours ago
- Business
- Wall Street Journal
Could Bringing AI Into the Physical World Make It Profitable? - What's News
As businesses are adopting artificial intelligence and beginning to figure out how it will make them money, developers are already working on ways to embody AI in the physical world. From home robots to manufacturing and beyond, tech reporter Belle Lin digs into the industry's plans and tells us whether physical AI might bring both makers and users the big returns on investment they've been anticipating. Alex Ossola hosts. Full Transcript This transcript was prepared by a transcription service. This version may not be in its final form and may be updated. Alex Ossola: Hey, What's News listeners? It's Sunday, June 15th. I'm Alex Ossola for the Wall Street Journal. This is What's News Sunday, the show where we tackle the big questions about the biggest stories in the news by reaching out to our colleagues across the newsroom to help explain what's happening in our world. On today's show, as businesses are finally starting to find ways to integrate artificial intelligence into their operations, developers are already working on future iterations of AI, including ways to embody the technology in the physical world. But the question remains can the developers or companies make money from AI? One of the biggest stories in tech over the past six months is the huge investments tech companies are making in data centers needed to power artificial intelligence. In January, Meta said it was allocating up to $65 billion this year, in the same month, Microsoft committed $80 billion, and in May a data center startup that works with OpenAI secured almost $12 billion. These developers have big plans. They see one of the next steps in artificial intelligence as bringing it out of the cloud and into the physical world like consumer devices and humanoid robots for manufacturing spaces. But will this future phase of AI finally earn a return on investment for these users and developers. To dig more into the AI industry's future plans and whether they'll make AI profitable, I'm joined by Belle Lin who covers AI and enterprise technology for the Journal. Belle, what do developers say is the next phase of AI? What's coming? Belle Lin: It's an interesting question because it feels like we're still in some of the earliest phases of AI where AI is still chatbots and you have to interact with ChatGPT in order to get something back, you have to type in something. But the wave after chatbots is supposed to be AI agents, and those are technologies or software that can basically do things for you, like order a cab when you're arriving home from the airport or to make a restaurant reservation. And then after that is physical AI and some tech watchers and certainly Jensen Huang, the CEO of Nvidia has talked about this phase as being where AI enters our physical world. And that has a lot of meanings, but in the corporate sense, it can mean that you're bringing automation to warehouses and bringing automation to factories. And then maybe in our daily lives, that's something like bringing humanoid robots to our homes. So broadly it's the idea that AI is entering our devices, whether in our homes, in wearable devices that we wear, or in the factories and the warehouses where our products and goods are made. Alex Ossola: I'm curious how that actually would work, because right now I think about AI as a chatbot essentially. How does that then become something that is embodied in the physical world, whatever that may mean? Belle Lin: There are some examples of wearable devices and these AI pins and devices that already came to fruition in this sort of first few phases of AI. There are things like AR and VR goggles that we've all heard of, the Apple Vision Pro. There's the Meta Quest, smart glasses, like from Meta and Snap. And so these are examples of AI that is embedded within these devices that we interact with, usually by voice or with gestures. Sometimes there's a more physical button that we might press or something that we might toggle, but the idea is really that AI gets embedded within the hardware itself rather than the human, the user, us being tied to some screen or some interface that we're used to seeing as a laptop or a phone. Alex Ossola: Who is leading this trajectory? Who's leading the pack? Belle Lin: What we've seen from OpenAI and Jony Ive's company is this collaboration called io, in which Jony Ive and his team will serve as the creative brain behind this new device that OpenAI will release, this sort of family of devices. And they've been pretty tight-lipped about what the device will look like and what it will do, but they've said a few things like it'll be ambient, it'll be this third core device that you put on your desk after your MacBook and your iPhone. And so you could say that they're leading the pack because they're promising a lot of what has yet to come, but they have this really great heritage in the whole Apple ecosystem and the design aesthetics that Jony Ive has put out. And also they have the models, they have the fantastic models that OpenAI has pioneered so far that are still state-of-the-art. So when you combine these two technology powerhouses right now, you get a bunch of promises, but they seem pretty promising. Alex Ossola: It sounds like there are a bunch of different kinds of applications, consumer-facing, more heavy industry, kind of something in between in the form of self-driving cars. Do we have a sense of which of these might sort of come first and how the developers of AI are thinking about monetizing those phases? Belle Lin: Monetization questions are always front and center because so many of these startups are funded by venture capital firms who need to see a return, and there's so much cash that's being injected into AI right now. Some of the ways in which they're monetizing are in the software side, on the models themselves. So you could sell on a word or a bit basis the ability to use OpenAI's models in other services and other technologies. In the wearable side, the selling of the hardware itself plus the software upgrades. But at this point, it's still really about adoption and figuring out which areas in the consumer world really stick. And then if we're talking about the heavy industry side, that's where ROI becomes a lot more important because you can shave a lot of costs by automating human labor away. And so that's where a lot of the warehouse and logistics companies are hoping to have an impact on their bottom lines. Alex Ossola: Coming up, AI developers may already be making the next generation of artificial intelligence, but if they build it, will the customers come? Stay with us. Belle, we've been talking a lot about the developer side, how AI gets made and what form it'll be in, but now I want to talk about the people who are going to be buying it and using it. Lots of companies have started using AI. According to a survey by McKinsey, 78% of companies say they use at least one AI function. So it seems like companies need to show they're integrating AI into their operations. Would you say this is an existential need for companies right now? Belle Lin: Oh, absolutely. There are really existential questions for categories of companies like law firms that have questioned what is the value of the billable hour, because so much of what AI is really good at automating away right now is reading and summarizing through texts and being able to provide synthesis of answers, and that's kind of early stage paralegal work. So if companies don't embrace AI, there's the question of will we still exist in 10 years timeframe? Never mind questions of will we be using AI pins and devices? We need to embrace AI now or else we won't be around. Alex Ossola: So that kind of brings me back to this other existential question about physical AI. Who actually wants this? Belle Lin: Well, if you look at examples of where physical AI exists now, I know we've talked about warehouses and factories. But there are also great examples of where wearable headsets like the Apple Vision Pro and the Meta Quest and many others that have been around for a while have huge applications in the military, for instance, for training the armed forces and in training for surgeries and home services where you have skilled trades like plumbers and air conditioning technicians, learning how to build the physical engines that keep homes running as well as jet engines, technicians learning and figuring out how to troubleshoot them. So there's great examples of where physical AI and augmented reality, which is a really early version of bringing AI into the real world, already have a lot of value. And so you might see more acceleration in areas where AI in the real world are already having an impact, but once it becomes much more useful, you could see things like basic knowledge work becoming a lot more augmented because the ability to stream someone's virtual presence into a meeting room makes it that much better and there's no longer a need to have an in-person meeting. Alex Ossola: One of the things that is in the news cycle about AI right now is just how unbelievably expensive it's been. Companies are shelling out billions of dollars to build these data centers. Because they are doubling down on AI being the future, is there enough demand in all of these different applications for physical AI that we've talked about that will bring down those costs of the data centers or will they just keep skyrocketing? Belle Lin: A lot of this goes back to the AI models and the software layer because as they become more efficient, then the promise is that they require a lot less GPU compute and power going into the data centers. And so when the models become more efficient themselves, even though they are quite large and unwieldy, they can be trained much more efficiently. From that point of view, costs will certainly start to come down in terms of the infrastructure. But at the same time, other costs will need to come down as well. The cost of hardware in a really general sense is still quite high, the chips required to basically power Apple Vision Pro or to power a humanoid robot or to power self-driving cars, those are not quite commoditized. They're still quite expensive. Alex Ossola: So as developers make these devices and software and as companies figure out how to use them, whose responsibility is it going to be to figure out how to actually make money off of this? Belle Lin: Yeah, a lot of the AI developers and the AI startups will be hard-pressed to come up with an answer on how to actually monetize what they're building. Right now, a lot of them are funded by VC dollars, are backed by research or other types of grants and funding. And so there will be this sort of inflection point where either their technologies or their devices, their robots, their cars catch on with consumers or they don't. Because as we look at some of the other waves of technology that were funded by VC dollars, like the Ubers and the Lyfts of the world, there's this limited timeframe in which they can be funded by venture capital dollars until they have to show their metal. Alex Ossola: And how about for the companies using the products? Belle Lin: For the companies, that's already a really pressing question. ROI has been challenging since the dawn of the ChatGPT, AI era that we're in now, about three years ago. Companies have been investing heavily in AI models and AI technologies, but there's really not a clear way to determine whether or not they're paying off. So you could say that productivity of workers has gone up, but it's hard to measure. You could say that sales have gone up, but that's also hard to measure. So measuring AI's value has been a question for tech executives for the past several years and continues to be, but there's a lot of economic incentives that are aligned in trying to make sure that the AI companies are profitable and that companies are saving on the bottom line and generating top line revenue that the market forces kind of end up working out in some way. Alex Ossola: That was WSJ reporter, Belle Lin. Thank you so much, Belle. Belle Lin: Thanks for having me. Alex Ossola: And that's it for What's New Sunday for June 15th. Today's show was produced by Charlotte Gartenberg with supervising producer Michael Kosmides and deputy editor Chris Zinsli. I'm Alex Ossola and we'll be back tomorrow morning with a brand new show. Until then, thanks for listening.
Forbes
3 days ago
- Business
- Forbes
What Is ‘Physical AI'? Inside The Push To Make AI Understand The Real World
What happens when AI enters the physical world — predicting actions, spotting risks and transforming ... More how machines understand real-time events? For years, AI has been great at seeing things. It can recognize faces, label objects and summarize the contents of a blurry image better than most humans. But ask it to explain why a person is pacing nervously near a fence, or predict what might happen next in a crowded room — and suddenly, the illusion of intelligence falls apart. Add to this reality the fact that AI largely remains a black box and engineers still struggle to explain why models behave erratically or how to correct them, and you might realize the big dilemma in the industry today. But that's where a growing wave of researchers and startups believe the next leap lies: not just in faster model training or flashier generative outputs, but in machines that truly understand the physical world — the way it moves, reacts and unfolds in real time. They're calling it 'physical AI'. The term was initially popularized by Nvidia CEO Jensen Huang, who previously has called physical AI the next AI wave, describing it as 'AI that understands the laws of physics,' moving beyond pixel labeling to bodily awareness — space, motion and interaction. At its core, physical AI merges computer vision, physics simulation and machine learning to teach machines cause and effect. Essentially, it enables AI systems to not just recognize objects or people, but to understand how they interact with their surroundings — like how a person's movement might cause a door to swing open or how a ball might bounce off a wall. At Lumana, a startup backed by global venture capital and growth equity firm Norwest, that phrase isn't just branding; it's a full-blown product shift. Known for AI video analytics, the company is now training its models not only to detect motion, but to recognize human behavior, interpret intent and automatically generate real-time alerts. 'We define physical AI as the next evolution of video intelligence,' Lumana CEO Sagi Ben-Moshe said in an interview. 'It's no longer just about identifying a red car or a person in a hallway — it's about inferring what might happen next, and taking meaningful action in real-world conditions.' In one real-world deployment, Lumana's system flagged a possible assault after detecting unusual body language and close proximity between two men and a pair of unattended drinks, prompting an alert that allowed staff to step in before anything escalated. In another case, it caught food safety violations in real time, including workers skipping handwashing, handling food without gloves and leaving raw ingredients out too long. These weren't issues discovered after the fact, but ones that the system caught as they unfolded. This kind of layered inference, Ben-Moshe explained, transforms cameras into 'intelligent sensors.' It's no coincidence that Huang has also previously used the term 'physical AI,' linking it to embodied intelligence and real-world simulation. It reflects a broader shift in the industry about creating AI systems that better understand the laws of physics and can reason more intelligently. Physics, in this context, is shorthand for cause and effect — the ability to reason about motion, force and interaction, not just appearances. That framing resonated with investors at Norwest, who incubated Lumana during its earliest phase. 'You can't build the future of video intelligence by just detecting objects,' said Dror Nahumi, a general partner at Norwest. 'You need systems that understand what's happening, in context and can do it better than a human watching a dozen screens. In many cases, businesses also need this information in real-time.' Norwest isn't alone. Other players, from Hakimo to Vintra, are exploring similar territory — using AI to spot safety violations in manufacturing, detect loitering in retail, or prevent public disturbances before they escalate. For example, Hakimo recently built an autonomous surveillance agent that prevented assaults, identified vandalism and even saved a collapsed individual using live video feeds and AI. At Nvidia GTC in March, Nvidia even demoed robotic agents learning to reason about gravity and spatial relationships directly from environment-based training, echoing the same physical reasoning that Lumana is building into its surveillance stack. And just yesterday, Meta announced the release of V- JEPA 2, 'a self-supervised foundation world model to understand physical reality, anticipate outcomes and plan efficient strategies.' As Michel Meyer, group product manager at the Core Learning and Reasoning arm of the company's Fundamental AI Research, noted on LinkedIn yesterday quoting Meta chief AI scientist Yann Lecun, 'this represents a fundamental shift toward AI systems that can reason, plan, and act through physical world models. To reach advanced machine intelligence, AI must go beyond perception and understand how the physical world works — anticipating dynamics, causality, and consequences. V‑JEPA 2 does just that.' When asked what the real-world impact of physical AI might look like, Nahumi noted that it's more than mere marketing. 'Anyone can detect motion, but if you want real AI in video surveillance, you must go beyond that to understand context.' He sees Lumana's full-stack, context-driven architecture as a foundation and not a vanity pitch. 'We think there's a big business here and the technology is now reliable enough to augment and outperform humans in real time,' told me. The reality is that the success of physical‑AI systems will not be just about the technology. As AI continues to advance, it's becoming much clearer that the success of most AI systems largely hinges on ethics, trust and accountability. Put in a different way, trust is the currency of AI success. And the big question that companies must continue to answer is: Can we trust your AI system to be safe? In a security context, false positives can shut down sites or wrongly accuse innocent people. In industrial settings, misinterpreted behavior could trigger unnecessary alarms. Privacy is another concern. While many physical AI systems operate on private premises — factories, campuses, hotels — critics warn that real-time behavior prediction, if left unchecked, could drift into mass surveillance. As Ben-Moshe himself acknowledged, this is powerful technology that must be used with guardrails, transparency and explicit consent. But, according to Nahumi, Lumana's multi-tiered model delivers actionable alerts, but also protects privacy and supports seamless integration into existing systems. 'Lumana engineers systems that layer physical AI on current infrastructure with minimal friction,' he noted, 'ensuring operators aren't overwhelmed by false positives.' Despite these questions, demand is accelerating. Retailers want to track foot traffic anomalies. Municipalities want to prevent crime without expanding staff. Manufacturers want safety compliance in real time, not post-event reviews. In every case, the challenge is the same: too many cameras, too little insight. And that's the business case behind physical AI. As Norwest's Nahumi put it, 'We're seeing clear ROI signals — not just in avoided losses, but in operational efficiency. This is no longer speculative deep tech. It's a platform bet.' That bet hinges on systems that are scalable, adaptable and cost-effective. Lumana's approach, which layers physical AI on top of existing camera infrastructure, avoids the 'rip-and-replace' problem and keeps adoption friction low. Nahumi pointed to rising enterprise demand across retail, manufacturing, hospitality and public safety — fields where video footage is ubiquitous, but analysis remains manual and inefficient. And even across boardrooms and labs, the appetite for machines that 'understand' rather than 'observe' is growing. That's why companies like Norwest, Nvidia, Hakimo and Lumana are doubling down on physical AI. 'In five years,' Ben-Moshe envisions, 'physical AI will do more than perceive — it will suggest actions, predict events and give safety teams unmatched visibility.' This, he noted, is about systems that not only see, but also act. Ultimately, the goal of physical AI isn't just to help machines see better — it's to help them understand what they're seeing. It's to help them perceive, understand and reason in the messy physical world we inhabit. Ben-Moshe envisions a future where physical AI suggests actions, prevents escalation and even predicts incidents before they unfold. 'Every second of video should generate insight,' he said. 'We want machines to reason about the world as a system — like particles tracing possible paths in physics — and highlight the most likely, most helpful outcome.' That's a far cry from today's basic surveillance. From thwarting crime and preventing accidents to uncovering new operational insights and analyzing activity trends, reasoning engines over cameras promise real, demonstrable value. But scaling them is where the real work is. It'll require systems that are accurate, ethical, auditable and trustworthy. If that balance is struck, we could enter a world where AI won't just help us see what happened, but help us know what matters most.
